Prior processing systems have included the technique of multiple users within a company sharing processing time available on a mainframe or central processing system. Using small segments of mainframe processing time, departments within the company would often incur costs associated with using the processing time, which in turn was billed back to each department from the central information technology (IT) organization for the company. In other instances, a company could pay for and utilize processing time made available by third-party companies who possessed an over-capacity of mainframe processing power. These third-party companies would, in effect, create a market for the mainframe processing time that went unused by the internal organizations of that third-party company.
Prior processing techniques have also included distributed processing projects that have utilized the Internet or World Wide Web. These distributed processing research projects have used personal computers (PCs) connected to the Internet to provide processing power to accomplish research project goals. Research project goals have been, for example, identifying large prime numbers, analyzing radio telescope data, and analyzing code keys in an encryption deciphering contest.
One example of a distributed processing project on the Internet is a research project housed at the University of California at Berkeley to analyze sky recording data gathered by SETI (the Search for Extraterrestrial Intelligence). This sky recording data has been gathered for some time from the large Arecibo Radio Telescope in Puerto Rico. The processing power needed to analyze these data recordings was very large. At the peak of SETI's capture activities, SETI had accumulated over 100,000 years of signals to process, as measured by the compute power necessary to process all the signals. To analyze this data, software was developed that could be downloaded to Internet connected PCs so that these PCs could process small slices of these sky recordings. In under a year, this project, called SETI@home (URL in March 2000—www.setiathome.ss1.berkeley.edu) has completely processed this backlog of data and is now returning to the sky recording dataset for further processing tasks. This massively parallel distributed system has a processing throughput of over 10 TFLOPs (terraFLOPS or 1012 floating point operations per second) running on about 1.8 million Internet connected machines.
Another example of a distributed processing technique was developed and implemented by Distributed.net (URL in March 2000—www.distributed.net) to compete in encryption breaking contests. Distributed.net created and distributed a client software program which may be downloaded by client systems connected to the Internet. This client software then acts as part of a large distributed processing system specifically designed to break encrypted messages on the Internet. Using this processing technique, Distributed.net has won encryption breaking contests sponsored by RSA Labs, which is an Internet security company. In these contests, RSA Labs has offered a monetary prize to the winner of the encryption contest. In organizing its efforts, Distributed.net has offered a share of this monetary prize to the client system that actually breaks the encryption code. In addition, Distributed.net keeps track of overall project statistics, as well as statistics concerning the efforts of its client systems through individual and team rankings by amount of processing completed.
Entropia.com (URL in March 2000—www.entropia.com) has utilized an Internet distributed processing system to compete in contests directed to identifying the largest prime number. Entropia.com also offers its computing power to other research projects. Users may sign on to be part of the distributed processing for free. For the largest prime number contest, Entropia.com, like Distributed.net, offers a monetary prize to the Internet connected PC that comes up with the first prime number achieved in a new order of magnitude. For other research projects, the incentive is simply to be a part of the research project.
Another distributing processing web site is provided by Process Tree Network (URL in March 2000—www.processtree.com). This web site is attempting to sign-up Internet connected computer systems to provide processing power for paying projects. For a project, each partner system, when connected to the Internet, will have client software that downloads a job unit and processes that job unit. The incentive offered by the Process Tree Network are “micro-payments” for the amount of work completed by any given system. These micro-payments are apparently small amounts of some total project value based upon the amount of the project completed by the given system through the jobs it has processed. In addition, each partner is given a bonus percentage of payments made to persons they sign-up as new partners.
In completely unrelated Internet activities outside the distributed processing arena, there have been a number of sites that have utilized a sweepstakes model as an incentive for consumer behavior. One of the most popular (as of March 2000) sweepstakes sites is IWON.COM (URL as of March 2000—www.iwon.com). IWON.COM is a standard Internet search and content portal that provides an incentive to users by giving them entries to a sweepstakes when the users use the portal. The more the users use the portal, the more entries the user generates, up to a limit, for example, up to 100/day. As of March 2000, at the end of each day, IWON.COM chooses a $10,000 winner from among the entries. At the end of each month, IWON.COM chooses a $1,000,000 winner. And, at the end of an overall sweeps period (as of March 2000), IWON.COM plans to draw a single winner for a $10,000,000 grand prize. IWON.COM has created this sweepstakes model to introduce an Internet portal in late 1999 and make it a web site that has as a comparable number of people using it as does Internet portals that have existed for many years, such as, for example, Yahoo.com (URL in March 2000—www.yahoo.com).
Significantly, these prior distributed processing projects have failed to fully utilize the capabilities of connected distributed devices.